Hemophilia A and B are X-linked recessive bleeding disorders deficient coagulation factor VIII (FVIII) and IX (FIX), respectively. Mutations in F8 and F9 gene are the genetic basis of hemophilia A and B. Affected males suffer from prolonged oozing after injuries or surgery and joint bleeding. The age at diagnosis and the severity of bleeding are closely related to the level of the residual coagulation factor activity [1, 2]. The diagnosis of hemophilia A and B is suspected based on the personal and family history of bleeding episodes and abnormal coagulation tests, which include prolonged activated partial thromboplastin time (aPTT) corrected on mixing test with normal plasma and decreased FVIII or FIX activity. Molecular genetic tests confirm the diagnosis in most patients with hemophilia A and B by identifying disease-causing mutations in F8 and F9, respectively [1, 2]. When a variation is detected, it is critical to gather further information and evidence to label it as a known or novel disease-causing mutation rather than a polymorphism. To date, 105 and 29 coding sequence variants of F8 and F9, respectively, are enlisted in the National Center for Biotechnology Information (NIH) single nucleotide polymorphism database (dbSNP) (last accessed March 20, 2013) [3]. Since the frequency and distribution of sequence variants vary across different populations, particularly in F8 [4], sequence variation data should be obtained from a panel of DNA samples of a control population from a given ethnic origin. In particular, variation data from control samples from functionally validated individuals are critical for reference in molecular diagnostics. In this study, we aimed to obtain sequence variation data of F8 and F9 in a set of control chromosomes from individuals of Korean descent and establish a set of sequence variation data to serve as a reference for the molecular diagnosis of hemophilia in Korean patients.

Thirteen individuals (13%) were shown to be hemizygous for a sequence variation in the coding sequence of F8 (Table 1). Six individuals (6%) had c.3780C>G (p.Asp1260Glu) [p.Asp1241Glu] and five individuals (5%) had c.3864A>C (p.Ser1288=) [p.Ser1269=]. One each individual had c.4794 G>T (p.Glu1598Asp) [p.Glu1579Asp] and c.5069A>G (p.Glu1690Gly) [p.Glu1671Gly] (Table 1). According to the 1000 Genomes database, c.3780C>G (Asp1260Glu) and c.3864A>C (Ser1288=) were known common SNPs (rs1800291 and rs1800292, respectively) with minor allele frequencies (MAF) of 24.6% and 10.5%, respectively. Since Asp1260Glu was previously described to be associated with the FVIII activity [4, 6, 7], we compared the FVIII activities of individuals carrying Asp1260Glu to those individuals without Asp1260Glu. The results showed that there was no significant difference in FVIII activities between the 2 groups by a univariate analysis (104% vs. 110% in individuals with Asp1260Glu and without Asp1260Glu, respectively, P=0.7883, Fig. 1). We also found no significant difference in FVIII activities between the carriers of Asp1260Glu and individuals without any variants (104% vs. 95%, P=0.8328). Glu1598Asp (Glu1579Asp) was enlisted in both HGMD and HAMSTeRS as a disease-causing mutation in a patient with severe hemophilia A, but not in dbSNP [8]. Glu1690Gly (Glu1671Gly) was a novel missense variation enlisted neither in public database nor in the literature [4, 5]. Comparative genomic analyses demonstrated neither Glu1598 nor Glu1690 residues were conserved among mammalian and non-mammalian species (Fig. 2). Functional predictions using SIFT and PolyPhen-2 demonstrated that Glu1598Asp was predicted to be tolerated and benign. Glu1690Gly was predicted to be tolerated by SIFT, but to be probably damaging by PolyPhen-2. Based on the normal FVIII activities in the control individuals carrying these variations (109% in individuals carrying Glu1598Asp and 148% in individuals carrying Glu1690Gly), both were considered to be rare SNPs rather than a causative mutation of hemophilia A. No sequence variation was observed in the F9 gene in the 100 control chromosomes.

Historically, linkage analysis was used to determine the mutation status in individuals with hemophilia [9-11]. In recent years, direct mutation detection techniques such as direct sequencing analysis, targeted mutation analysis (intron 22 and intron 1 inversion mutations of F8), and deletion/duplication analysis have replaced linkage analysis, and these direct techniques have detected disease-causing mutations in >95% of patients with hemophilia [12-16]. Given the scarcity of mutation hotspots within these genes and the large size of F8, a significant proportion of the hemophilia mutations in recent reports are still novel, and it is critical to determine whether the novel variation is a disease causing mutation or polymorphism. In particular, when the variation is neither apparently deleterious nor silent but is predicted to change the amino acid, deciding whether the variation is a disease-causing mutation or a polymorphism becomes highly challenging. For these reasons, variation dataset from functionally validated (factor levels within reference ranges) control samples are crucial for molecular diagnosis of hemophilia. In particular, male control samples from ethnic origin of interest are relevant to hemophilia diagnostics because carrier females of hemophilia A/B may have a normal factor level and the SNP frequencies throughout the entire F8 gene differ widely across different ethnic groups [4, 17]. According to the validation of mutations approved by the Human Genome Organization (HUGO) Mutation Database Initiative/Human Genome Variation Society (HGVS), the determination of disease-causativeness needs to be supported by observation of 0% frequency of the variation in 100 normal chromosomes.

In the present study, we obtained a set of variation data for the F8 gene in 100 control male individuals (100 X chromosomes) of Korean descent with normal FVIII activities. A significant proportion (13%; 13/100) of the control chromosomes carried sequence variations in F8. Asp1260Glu (Asp1241Glu) and Ser1288= (Ser1269=) are common SNPs previously enlisted in the dbSNP (rs1800291 and rs1800292, respectively). The frequencies of both Asp1260Glu and Ser1288= in Korean were lower than the minor allele frequencies in the dbSNP (6% vs. 24.5%, 5% vs. 10.5%, respectively). c.3780C>G (p.Asp1260Glu) demonstrated a large difference in the frequency of the variant G allele. The frequency of the variant G allele was reported at 3.4% in Japanese, 7.8% in Han Chinese, 15.3% in European, and 69.2% in Sub-Saharan African individuals [3]. Thus, the frequency of c.3780C>G (Asp1260Glu) in Korean control individuals (6%) was between those reported for Japanese and Han Chinese individuals. c.3864A>C (p.Ser1288=) is a common SNP across population and the frequency of the variant C allele in this study was 5%, which fell between those reported for Japanese (2.2%) and Han Chinese (7.8%) individuals [3]. The frequency of the variant allele in the Korean individuals was similar to that of sub-Saharan African individuals (5.8%) and lower than that of European individuals (8.3%). The aspartate residue at 1260 is located in the B domain of the FVIII protein, and Asp1260Glu was reported to be associated with decreased FVIII activity (by .10%) [4, 6, 7]. However, the difference of FVIII activities was not significant between individuals with or without Asp1260Glu in our study, possibly due to the limited number of samples or other factors such as age, ABO blood type, smoking, and plasma level of the von Willebrand factor [4]. Glu1598Asp (Glu1579Asp) was registered in LSDB (HAMSTeRS) and HGMD as a disease-causing mutation in a patient with severe hemophilia A, but was not registered in dbSNP and other polymorphism databases [8]. However, our data indicate that Glu1598Asp is a rare SNP rather than a mutation associated with hemophilia A because it was found in our control chromosomes. Glu1690Gly (Glu1671Gly) was a novel amino acid-changing variation and was found in 1%(1/100) of control chromosomes. Based on the normal FVIII activity in control individual carrying this variation (148%), it was considered a rare SNP, also. Unlike F8, no variation was observed in coding sequences of F9 in all 100 of the control chromosomes. Although our control sample set was not tested for FIX activity, the result implicates that F9 is less polymorphic in Korean descendants compared to F8. So, when a novel amino acid-changing variation is found in the F9 gene in a patient suspected to have hemophilia B, one should search for other evidence to prove disease causativeness of the variation.

The results obtained in this study can be used as a reference dataset for molecular diagnosis of hemophilia A and B in Korea. In addition, the data we obtained revealed rare SNPs, including one enlisted as a missense mutation causing hemophilia A. The data presented here indicate that interpretation of sequencing data using public databases to diagnose hemophilia A and B should be approached cautiously. It should also be noted that the number of chromosomes (100) used in this study has a limitation in identifying other rare SNPs, especially those with a minor allele frequency <1% [18].